Observationally constrained values of the global radiative response coefficient are pivotal to assess the reliability of modeled climate feedbacks. A widely used approach is to measure transient global radiative imbalance related to surface temperature changes. However, in this approach, a potential error in the estimate of radiative response coefficients may arise from surface inhomogeneity in the climate system. We examined this issue theoretically using a simple two-zone energy balance model. Here, we dealt with the potential error by subtracting the prescribed radiative response coefficient from those calculated within the two-zone framework. Each zone was characterized by the different magnitude of the radiative response coefficient and the surface heat capacity, and the dynamical heat transport in the atmosphere between the zones was parameterized as a linear function of the temperature difference between the zones. Then, the model system was forced by randomly generated monthly varying forcing mimicking time-varying forcing like an observation. The repeated simulations showed that inhomogeneous surface heat capacity causes considerable miscalculation (down to −1.4 W m−2 K−1 equivalent to 31.3% of the prescribed value) in the global radiative response coefficient. Also, the dynamical heat transport reduced this miscalculation driven by inhomogeneity of surface heat capacity. Therefore, the estimation of radiative response coefficients using the surface temperature-radiation relation is appropriate for homogeneous surface areas least affected by the exterior.